Labor Market Integration of Refugees: RCT Evidence from Sweden

How to cite

Dahlberg, Matz; Vikman, Ulrika; Egerbark, Johan. 2019. "Labor Market Integration of Refugees: RCT Evidence from Sweden." AEA RCT Registry. AEARCTR-0004788. https://doi.org/10.1257/rct.4788-1.0

Download: BibTeX · RIS — for reference managers without a browser translator. Zotero and most others pick up this page's citation metadata directly.

Matz Dahlberg* , Ulrika Vikman , Johan Egerbark

* Primary investigator

Registry ID
AEARCTR-0004788
DOI
https://doi.org/10.1257/rct.4788-1.0
First registered
Last updated
Status
On going
Trial dates
2017-03-01 – 2020-06-30
Countries
Private
JEL codes
J08, J15, J16
Registry record
http://www.socialscienceregistry.org/trials/4788

Abstract

The recent large inflows of refugees to many Western countries has spurred a large, and often intense, debate on how to best integrate the refugees into their new countries. Finding successful integration policies is very important from both the refugees' and the host countries' perspective.

In this project we will evaluate a new and uniquely ambitious integration program that was developed and put into effect in Sweden. The aim with the program, run in cooperation by the City of Gothenburg and the Public Employment Office (PES) in Gothenburg, was to get newly arrived refugees to Sweden better integrated (in terms of, e.g., labor market outcomes).

The evaluation will be based on a randomized experiment, in which potential participants in the program in one of the waves were randomly assigned into treatment and control groups.

Intervention

When starting the new type of very ambitious integration program, run in cooperation by the City of Gothenburg in Sweden and the Public Employment Office (PES) in Gothenburg, with the aim to get newly arrived refugees to Sweden better integrated (in terms of, e.g., labor market outcomes), it was decided that there should be 50 slots available in the program for each starting cohort. Our experiment is for the second starting cohort (the starting date for this cohort was April 20, 2017, when prospective participants were called to an information meeting about the program at the Public Employment Service (PES) in Gothenburg. The program started on May 29, 2017.) We use a lottery to decide which of all eligible individuals that should be given the opportunity to participate in the new program.

Experimental design

Given the eligibility criteria for participating in the program (based on the refugees' educational level and time in Sweden), 140 individuals were identified as being potential participants in the program. The eligibility criteria was stated by one of the two organizations running the program, the City of Gothenburg, and the identification of eligible candidates was made by the other responsible organization, the Public Employment Service in Gothenburg.

Out of the 140 potential participants, 70 individuals were randomly drawn to be called to an information meeting about the program. Since we know since earlier that labor market outcomes are strongly correlated with (immigrants') gender in Sweden, we block-randomize based on gender (the randomization make the share of men and women in the treatment group to match their shares in the population). The randomization was made by the analytical unit at the Swedish Public Employment Service in Stockholm. At the information meeting, they were given an explanation of the program and was given an offer to participate in the program. 50 individuals accepted the offer to start in the new program. The 70 individuals that were selected by the lottery to be given an opportunity to participate in the program are all defined to be in the treatment group. We are hence estimating Intention-to-Treat (ITT) parameters.

The 70 eligible individuals that, by randomization, was selected to not participate in the program constitute the control group. They participate in the traditional programs run by the PES in Gothenburg.

In a first step, we will evaluate the short-run effects of the new program on labor market outcomes. Short-run is here meant to be within two years; we will evaluate the effects at different time points within these two years. We take the start of the program on May 29, 2017 as the starting point for the program and May 29, 2019, as the end point for the two year evaluation. The outcome variables are measured via the registers held by the PES in Sweden. We will also evaluate the effects on intermediate outcomes such as native language development and adult education in Sweden.

Design details

Randomization method
The randomization was made by the analytical unit at the Swedish Public Employment Service in Stockholm. The random draws were made in a computer (the program was written in Stata).
Randomization unit
Individuals, stratified on gender
Clusters
2 clusters, males and females
Observations
140 individuals
Arms
70 individuals in treatment group, 70 individuals in control group
Primary outcomes
Labor market outcomes (time to employment, type of employment)
Secondary outcomes
Native language development (results on language tests) and adult education (type of education and results) in Sweden.

Cited by

No citation data available for this registration yet.

Keywords